We agree with Dr. Weil that the changes in biochemical markers reflecting bone turnover are not striking. We have used lower doses of fluoride than did previous trials that showed quicker and higher increases in bone formation, evidenced by an increase in biochemical markers of bone formation and an increase in bone mineral density (mainly at trabecular sites). Nevertheless, the increase in serum alkaline phosphatase level reached the level of statistical significance at the end of our study. Given such a low dose of fluoride, the size of our sample and the high variability of biochemical determinations of markers of bone remodeling, it would have been surprising to see a statistically significant difference in these variables earlier in the trial. The difference in the mean bone specific serum alkaline phosphatase between the fluoride plus calcium group and the calcium-only group is in the range of 15% after 6 months and 20% after 12 months. The urine hydroxyproline-to-creatinine ratio was the only marker available to us at the time of initiation of this protocol (6 years ago). The new biochemical markers of bone resorption might have been much more sensible and could have provided much more valuable information. Furthermore, the predictive value of biochemical markers of bone remodeling as a surrogate for changes in bone mineral density has been validated only for inhibitors of bone resorption (for example, estrogens and bisphosphonates). So far, no studies have found a clear correlation between changes in markers and later changes in bone mineral density for stimulators of bone formation.

Regarding Dr. Weil's second comment, our sample consisted largely of women relatively far past menopause. Few data exist about the use of fluoride in women immediately after menopause.

We are afraid that Dr. Lesser's interpretation of our data relates to a gross misunderstanding of our Discussion. Our sample was relatively small, and few fractures occurred during our study. Nevertheless, twice the number of fractures (2/1) occurred in the calcium-only group compared with the fluoride plus calcium group after the second year of treatment. Furthermore, our Discussion was based on an extensive reading of the FAVOS report, on which we were one of the major contributors. That study considered a much bigger sample and had more fracture events. In this particular study, fractures were equally distributed between the two groups during the 2 years of treatment, and no trend toward reduction in vertebral fractures under fluoride treatment was observed. In our paper, the difference between the two groups at the second year was not statistically significant, notwithstanding, as Dr. Lesser noticed, more fractures in the calcium-only group during the first year of the trial. The statistical difference was reached only after 4 years. This confirms the statistical importance of the fracture events occurring during the third and fourth years. Therefore, even if we cannot totally rule out a certain oddity in the occurrence of fractures during the first year of the study, we can only confirm that this potential oddity was not accompanied by any statistical significance. Thus, we stand by the conclusion of our paper.